Alibaba’s Qwen3.7-Plus supports text, video and imagery inputs at low cost of $0.4/$1.6 per 1M token — but it’s proprietary
Publish Date: 2026-06-02 18:40:00
Source Domain: venturebeat.com
-
New AI Model Release: Alibaba launched Qwen3.7-Plus, the latest in its Qwen family of AI large language models, offering more multimodal capabilities and a 60% lower cost than its previous text-only model, Qwen3.7-Max.
-
Closed-Access Strategy: Unlike previous open-source Qwen models, Qwen3.7-Plus is now available only under a closed commercial license through proprietary APIs. This marks a shift for companies like Airbnb that relied on open-source versions.
-
Highlighted Features: The model stands out with low cost and superior performance in multimodal tasks such as creating enterprise-grade visuals, analyzing video, imagery, and screenshots—functions Qwen3.7-Max lacks as it’s text-only.
-
Competitive Pricing: Qwen3.7-Plus is positioned as one of the cheaper powerful AI models with competitive pricing just marginally above its Chinese rival’s discounted offering.
-
Architectural Improvements: The model offers improved capabilities for managing context and preserving reasoning state through its ‘preserve_thinking’ parameter, aiming to address state decay in long-horizon tasks typically executed by autonomous agents.
-
Benchmark Performance: While it delivers competitive results on various benchmarks, Qwen3.7-Plus falls below state-of-the-art models from competitors like Anthropic and OpenAI.
-
Target Use Cases: Best suited as a cost-effective replacement for premier frontier models in high-frequency workflows, Qwen3.7-Plus integrates easily with existing frameworks and offers significant cost reductions for repetitive tasks through caching.
-
Commercial Licensing and Compliance: Unlike previous Qwen models which were fully open-source, Qwen3.7-Plus only operates under a managed commercial API framework, raising concerns regarding compliance, data residency, and lack of on-premise deployment options.
